Towards a Superorganismal Perspective of Honey Bee Microbiota: Taxonomic Resolution by Niche and 16S rRNA Gene Variable Region

Abstract The honey bee ( Apis mellifera ) colony is a superorganism, but its microbiota is known mostly from the worker hindgut. Direct comparison across studies and niches is confounded by different regions of the 16S rRNA gene, bioinformatic processing choices, and reference database used for taxonomic assignment. We reprocessed 4,586 short-read Illumina libraries from 37 BioProjects, and added 345 full-length (V1-V9) PacBio libraries of dissected worker and queen gut tissues as a reference. Sequences were processed using one set of filtering and reporting thresholds and one species-level reference database. Each library was processed twice, with and without read denoising. The two methods agreed on genus-level composition but denoised reads produced far fewer amplicon sequence variants per library and significantly lower alpha diversity. Species resolution was highly dependent on 16​S rRNA gene region. Full-length reads named nearly every major genus to species, while some short regions named almost none. The superorganismal microbiota is comprised of both niche specialists, and generalists shared across multiple niches. We identified at least three abundant amplicon sequencing variants that lack species-level description. Contaminant reads were abundant throughout the dataset due to next generation sequencing of niches with low microbial biomass. Contaminants scaled inversely with bacterial load, dominating drone reproductive tissue, larvae, mouthparts, hive materials, control or lab reared worker guts and stored food. Thus, conclusions about honey bee microbiota and/or its integration with host information depend strongly on primer region, how sequencing error was handled, and the level of high-throughput sequencing contamination.